Abstract

Competing theories of why voters in Weimar Germany chose the NSDAP (Nazi party) have been only partially supported; the notion of a “catch-all” party, with voters in different classes and social groups, is gaining adherents. Previous research has treated the German electrorate as a national unit. Regional and local elements of the voting surface have been generally ignored, even though strong historical and material conditions generated important deviations from the national trends. Using descriptive spatial statistics, this article documents the importance of spatial heterogeneity and dependence in the 1930 Nazi vote. A mixed structural-spatial model, in whichkey variables from the several theoretical explanations of the NSDAP vote are included with geographic variables, demonstrates the importance of spatial and contextual effects. Regional variations from the average NSDAP vote (18.3 percent in 1930) persist even after religious and class effects are controlled. Accordingly, domain-specific models based on the regions of Weimar Germany are preferable to national models. The former indicate that specific combinations and relative significances of the explanatory factors vary from region to region. Domain-specific models sustain electoral geography's central tenet, namely, that places and contexts influence voting choices in addition to the social characteristics of the voter. Context introduces a new and important element in the interpretation of the Nazi rise to power.